| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 24 | | adverbTagCount | 1 | | adverbTags | | 0 | "She pauses then [then]" |
| | dialogueSentences | 61 | | tagDensity | 0.393 | | leniency | 0.787 | | rawRatio | 0.042 | | effectiveRatio | 0.033 | |
| 96.80% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1563 | | totalAiIsmAdverbs | 1 | | found | | | highlights | | |
| 100.00% | AI-ism character names | Target: 0 AI-default names (16 tracked, −20% each) | | codexExemptions | | | found | (empty) | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 93.60% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1563 | | totalAiIsms | 2 | | found | | | highlights | | |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 0 | | maxInWindow | 0 | | found | (empty) | | highlights | (empty) | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 0 | | narrationSentences | 123 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 1 | | hedgeCount | 1 | | narrationSentences | 123 | | filterMatches | | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 154 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 51 | | ratio | 0 | | matches | (empty) | |
| 66.15% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 30 | | markdownWords | 104 | | totalWords | 1554 | | ratio | 0.067 | | matches | | 0 | "fresh" | | 1 | "“You’ll never be worth saving,”" | | 2 | "“Not even to yourself.”" | | 3 | "Eva." | | 4 | "Eva" | | 5 | "“You’ll rot, Laila—you’ll rot like your mother.”" | | 6 | "The Raven’s Nest" | | 7 | "Laila" | | 8 | "“We’re Laila and Aurora now,”" | | 9 | "“No one gets to hurt Laila but me.”" | | 10 | "“And me,”" | | 11 | "“I’ll hurt anyone who hurts Laila.”" | | 12 | "same" | | 13 | "stubborn" | | 14 | "her" | | 15 | "To Kill a Mockingbird" | | 16 | "“It’s okay, Mr. Whiskers,”" | | 17 | "“I won’t let anyone take you away.”" | | 18 | "“You can’t even take care of yourself, let alone a cat.”" | | 19 | "“Try me,”" | | 20 | "“I’ll break your fingers.”" | | 21 | "“You should’ve called,”" | | 22 | "“We’re worried. About you. About Laila.”" | | 23 | "“I’m not Laila anymore, Mom,”" | | 24 | "“I’m Rory. Aurora. Whatever.”" | | 25 | "named" | | 26 | "suffer" | | 27 | "Deliveries" | | 28 | "“You’ll rot, Laila,”" | | 29 | "hiding" |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 30 | | unquotedAttributions | 0 | | matches | (empty) | |
| 0.00% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 69 | | wordCount | 899 | | uniqueNames | 10 | | maxNameDensity | 2.78 | | worstName | "Eva" | | maxWindowNameDensity | 6 | | worstWindowName | "Eva" | | discoveredNames | | Soho | 1 | | Evan | 8 | | Rory | 22 | | Cardiff | 1 | | Raven | 3 | | London | 2 | | Eva | 25 | | Silas | 3 | | Camden | 1 | | Nest | 3 |
| | persons | | 0 | "Evan" | | 1 | "Rory" | | 2 | "Raven" | | 3 | "Eva" | | 4 | "Silas" | | 5 | "Nest" |
| | places | | 0 | "Soho" | | 1 | "Cardiff" | | 2 | "London" | | 3 | "Camden" |
| | globalScore | 0.11 | | windowScore | 0 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 63 | | glossingSentenceCount | 0 | | matches | (empty) | |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 1 | | per1kWords | 0.644 | | wordCount | 1554 | | matches | | 0 | "No one gets to hurt Laila but" |
| |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 1 | | totalSentences | 154 | | matches | | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 49 | | mean | 31.71 | | std | 22.64 | | cv | 0.714 | | sampleLengths | | 0 | 57 | | 1 | 68 | | 2 | 83 | | 3 | 23 | | 4 | 35 | | 5 | 52 | | 6 | 10 | | 7 | 52 | | 8 | 29 | | 9 | 11 | | 10 | 59 | | 11 | 16 | | 12 | 5 | | 13 | 13 | | 14 | 9 | | 15 | 49 | | 16 | 14 | | 17 | 60 | | 18 | 18 | | 19 | 53 | | 20 | 59 | | 21 | 27 | | 22 | 81 | | 23 | 18 | | 24 | 6 | | 25 | 69 | | 26 | 40 | | 27 | 6 | | 28 | 24 | | 29 | 42 | | 30 | 5 | | 31 | 59 | | 32 | 17 | | 33 | 52 | | 34 | 50 | | 35 | 26 | | 36 | 18 | | 37 | 18 | | 38 | 50 | | 39 | 45 | | 40 | 9 | | 41 | 27 | | 42 | 4 | | 43 | 36 | | 44 | 6 | | 45 | 34 | | 46 | 5 | | 47 | 2 | | 48 | 3 |
| |
| 99.56% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 2 | | totalSentences | 123 | | matches | | |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 160 | | matches | (empty) | |
| 68.65% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 5 | | semicolonCount | 0 | | flaggedSentences | 4 | | totalSentences | 154 | | ratio | 0.026 | | matches | | 0 | "Most nights, the bar hums like a well-oiled engine—drunks slurring, a jazz record crackling from the speakers, the clatter of a glass that’s been set down too hard." | | 1 | "Rory’s hand drifts to her left wrist, where the scar still itches—thin, white, a crescent from the day she tried to slash her way out of Evan’s car, 10 years ago." | | 2 | "The neon green sign—*The Raven’s Nest*—paints them both in sickly light." | | 3 | "“He’s got a meeting in the back. The one with the bookshelf. Told me to wait.” She pauses, glancing at the maps on the wall—old, crumpled, pins marking London." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 904 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 28 | | adverbRatio | 0.030973451327433628 | | lyAdverbCount | 5 | | lyAdverbRatio | 0.0055309734513274336 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 154 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 154 | | mean | 10.09 | | std | 9.49 | | cv | 0.94 | | sampleLengths | | 0 | 6 | | 1 | 10 | | 2 | 9 | | 3 | 2 | | 4 | 28 | | 5 | 2 | | 6 | 5 | | 7 | 7 | | 8 | 14 | | 9 | 31 | | 10 | 7 | | 11 | 4 | | 12 | 4 | | 13 | 26 | | 14 | 4 | | 15 | 17 | | 16 | 32 | | 17 | 11 | | 18 | 12 | | 19 | 2 | | 20 | 8 | | 21 | 12 | | 22 | 4 | | 23 | 9 | | 24 | 6 | | 25 | 11 | | 26 | 7 | | 27 | 14 | | 28 | 14 | | 29 | 4 | | 30 | 4 | | 31 | 2 | | 32 | 9 | | 33 | 11 | | 34 | 29 | | 35 | 3 | | 36 | 6 | | 37 | 13 | | 38 | 10 | | 39 | 6 | | 40 | 5 | | 41 | 3 | | 42 | 1 | | 43 | 19 | | 44 | 17 | | 45 | 13 | | 46 | 6 | | 47 | 3 | | 48 | 13 | | 49 | 3 |
| |
| 51.08% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 3 | | diversityRatio | 0.3181818181818182 | | totalSentences | 154 | | uniqueOpeners | 49 | |
| 32.36% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 1 | | totalSentences | 103 | | matches | | 0 | "Somewhere, a siren wails." |
| | ratio | 0.01 | |
| 87.96% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 34 | | totalSentences | 103 | | matches | | 0 | "She straightens, her apron crinkling." | | 1 | "Her best friend in Cardiff," | | 2 | "It’s a statement, like she’s" | | 3 | "Her fingers close around a" | | 4 | "She sets it down, careful," | | 5 | "It comes out flat, like" | | 6 | "She’s wearing a charcoal suit," | | 7 | "She nods, too small." | | 8 | "She steps past the bar," | | 9 | "She pauses, glancing at the" | | 10 | "She grabs two tumblers, fills" | | 11 | "They used to only ever" | | 12 | "she’d whispered once, as they’d" | | 13 | "She leans on the bar," | | 14 | "She pauses, then" | | 15 | "She slides down onto a" | | 16 | "She can see it: Eva," | | 17 | "She opens her eyes." | | 18 | "She flexes her fingers, the" | | 19 | "She pulls out a file," |
| | ratio | 0.33 | |
| 23.11% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 90 | | totalSentences | 103 | | matches | | 0 | "The bell above the door" | | 1 | "Rory glances up from where" | | 2 | "The sound’s too sharp, too" | | 3 | "She straightens, her apron crinkling." | | 4 | "The figure in the doorway" | | 5 | "The light catches a silver" | | 6 | "Rory’s hand drifts to her" | | 7 | "The figure steps closer." | | 8 | "*Eva.* But not *Eva*." | | 9 | "Her best friend in Cardiff," | | 10 | "The voice is a question," | | 11 | "It’s a statement, like she’s" | | 12 | "Her fingers close around a" | | 13 | "She sets it down, careful," | | 14 | "It comes out flat, like" | | 15 | "Eva pushes the door open" | | 16 | "The neon green sign—*The Raven’s" | | 17 | "She’s wearing a charcoal suit," | | 18 | "Rory’s throat feels tight." | | 19 | "She nods, too small." |
| | ratio | 0.874 | |
| 97.09% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 2 | | totalSentences | 103 | | matches | | 0 | "Now Rory sees the gray" | | 1 | "Now Eva takes the glass," |
| | ratio | 0.019 | |
| 91.84% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 28 | | technicalSentenceCount | 2 | | matches | | 0 | "Most nights, the bar hums like a well-oiled engine—drunks slurring, a jazz record crackling from the speakers, the clatter of a glass that’s been set down too h…" | | 1 | "Her best friend in Cardiff, who’d cried with her over takeaway curries when Evan first shoved her." |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 24 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 7 | | fancyCount | 3 | | fancyTags | | 0 | "the police came (come)" | | 1 | "she’d (would)" | | 2 | "her mother had (have)" |
| | dialogueSentences | 61 | | tagDensity | 0.115 | | leniency | 0.23 | | rawRatio | 0.429 | | effectiveRatio | 0.098 | |